EIS Arapahoe & Roosevelt National Forest, Pawnee National Grassland

Home Browse Search

30°l0 ot' the basal ai•ea would be expected to produce less than 30% of the water yield of a clear cut). This has not been borne out by more recent research. ? The FORPLAN model for the Arapaho-Roosevelt produced a harvest categoiy called "special cuts". This was primarily harvest that would be done to benetit other functions, slthough it would be accomplished through commercial timber haivest. After consultation with the forest silviculturist and wildlife biologist about the kind of treatments which could be expected. I treated lodgepole pine special cuts as clearcuts. spruce-tir special cuts as first step shelteiwood treatments, and ponderosa pine special cuts as if they would produce no water yield increase because of the lightness of the treatment. How special cuts are treated in estimating water yield increases may vary from Forest to Forest. Water yield should only be modeled for the first tive decades. Regulations require that the FORPLAN model be run for an entire rotation, but that is primarily to ensure that ASQ is sustainable. I think that five decades is long enough to show trends in water yield increases from proposed harvest, and it is also the time period used for analysis of most other resources. I had initially suggested that basal area be estimated by simply averaging the basal area of the suitable-scheduled stands. Rudy King (RMS biometrician) pointed out that the relationship of water yield to change in basal area was not linear, and that averaging the basal areas to produce an average water yield was incorrect. He suggested that predicting water yield for a number of classes of basal areas and averaging the results would be more robust. After examining the results of the FVS modeling and experimenting with WRENSS, it appears that water yield is relatively insensitive to the limited range of pre-treatment basal areas predicted by FVS for the first five decades. Because of this, I propose that we model water yield for the smallest and greatest basal areas predicted by FVS for the five decades of treatment. If there is no change in water yield, there is no need to look at inteimediate conditions. If there is a change. the intermediate basal areas should also be modeled. The following is a block by block discussion ot the data entiy required to drive WRENSS. Area - I arbitrarily selected 100 acres. Because the model delivers results in depth, rather than volume, and because we desire unit water yield, area is not pazticularly important. However, I wanted a unit large enough so that I didn't experience the logical inconsistency of having windward widths or harvest block areas that were bigger than the unit areas (although it doesn't appear to matter to the model). Harvest of the entire unit should be simulated. Aspect - Each combination of vegetation type -treatment- location should be modeled for each aspect. Because aspect is a discrete rather than a continuos variable, you can't calculate a weighted mean average. The model is also very sesitive to aspect. I suggest modeling water yield for each aspect and then calculating a weighted mean water yield based on the area of a vegetation type occupying each aspect (this information can be obtained from GIS or the RIS database). This is explained in greater detail in the section following data entry. Precipitation - Because of the general dearth of precipitation data on this, and most other Forests. I think each Forest needs to be given considerable latitude in using the data which is available. This is the method I used. I digitized the isohyetal lines from the "Colorado Average Annual Precipitation Map, 1951-1980" (Colorado Climate Center, 1984) into our GIS. I then J 5